ANATOMY AND PHYSIOLOGY OF ATRINA RIGIDA. 
419 
a slight pull is necessary to separate them. In this way the branchial and suprabran- 
chial chambers can be thrown together. In fact, the animal can maintain them separate 
or throw them together apparently at will. These unions commonly show an inter- 
locking of the epithelial cells as well as the cilia, and sometimes the epithelium of the 
opposed surfaces is thrown into a series of ridges and grooves, thus producing a firmer 
union. Although Lacaze-Duthiers (8) and Peck (12) have described forms in which 
there is a weak union between the upper borders of the reflexed lamellae and the mantle, 
they did not describe the actual mode of union. Their purpose in the description was to 
show the transition between those forms which have the mantle edge free and those 
which have it permanently united to the neighboring parts. Grobben (5) has shown 
that this weak union is by means of interlocking cilia, and he considers it to be universal 
among the Aviculidae. He states also that when the opposed surfaces are forcibly 
separated they will reunite in a short time if undisturbed. The evidence upon which 
this assumption was based was his observations on the gill of Mytilus. Here he found, 
what Lacaze-Duthiers had already described, that the ciliary interfilamentar connectives 
would reform after being separated. Stenta (14) demonstrated that the reflexed lamellae 
of the gills of Pinna and Solen would reunite to adjacent parts after being separated 
from them, and I have confirmed the same for Atrina. I separated the gill from its 
attachment to the mantle for a distance of 2 inches. When examined several hours 
later it had effected a union. Stenta thinks that this type of union between the gills 
and mantle is of much more general occurrence than has been supposed, suggesting 
that it may occur in those forms in which the gills have been described as free. He 
maintains that in life they are never separated unless by accident, but he is probably in 
error, because I observed the mantle gland, which normally lies in the cloacal chamber, 
extending far down into the branchial chamber. This could not take place so long as the 
gills retained their connection with each other. 
When magnified sufficiently each lamella is seen to be thrown into a series of folds 
(grooves and ridges). These structures are barely visible to the unaided eye as a series 
of parallel lines running across the gill perpendicular to its base. Each ridge (fig. 3) 
is composed of from 10 to 12 hollow filaments which are slightly separated from each 
other. The latter are bound together at regular intervals by tubular interfilamentar 
connectives which are somewhat larger than the filaments and run at right angles to 
them. These two sets of tubules thus form a trelliswork in which the spaces between 
are the ostea through which water enters the gill from the branchial chamber. (Fig. 4.) 
The one or two filaments which occupy the summit of the ridge differ somewhat from 
the others in that they contain numerous goblet cells whose sticky secretion entangles 
minute organisms as they are carried over the gills in the respiratory current. The fila- 
ments and ridges of one lamella do not pass directly over into those of the other, but 
gradually decrease in size and disappear as they approach the edge of the gill. There 
is a deep groove with smooth walls in the edge of each gill which is lined by ciliated 
epithelium. (Fig. 2, g.) There is no fusion of filaments due to crowding as has been 
described by Rice for Cardium and other forms (13). 
